RCC坝力学参数反演不唯一性概率统计分析方法探讨

由于室内试验确定的混凝土坝力学参数与实际参数存在较大差异,目前混凝土坝工程上常基于实测变形采用优化算法或仿生算法反演获得混凝土坝力学参数,然而多参数反演不唯一性问题尚未解决。针对反分析不唯一性问题,考虑到大坝混凝土弹性模量和强度之间密切相关,基于大坝混凝土强度标准值的定义方法,建议基于实测变形进行多次力学参数反演,然后对反演结果进行概率统计分析获得概率分布函数,依据80%保证率确定反演参数。结合高寒地区某碾压混凝土(RCC)坝实测变形,验证了本文提出的反演参数不唯一性概率统计分析方法。分析表明该方法可以获得相对稳定的坝体及坝基力学参数反演值,可为高寒地区碾压混凝土坝安全性态评估提供参考。

A Probability Statistical Analysis Method for Nonuniqueness of Mechanical Parameters Inversion of Roller Compacted Concrete Dam

As the mechanical parameters of concrete dam obtained by indoor test deviates largely from measured values, back analysis of parameters using optimization or bionic algorithm based on measured deformation is adopted in practical engineering. To address the nonuniqueness of parameter inversion, we propose to obtain the probability distribution function of multiple back analysis results based on measured deformation via probability statistical approach, and then determine the inversion parameters according to 80% guarantee rate. The proposed method is verified by the measured deformation values of a roller compacted concrete (RCC) dam in high and cold region. The method provides stable mechanical parameters of dam body and dam foundation, and offers a reference for the safety assessment of RCC dam in high and cold region.